Ranging from disdain to contempt, the dissatisfaction with DSM and ICD categories takes many forms. Some reject the value of official, shared classifications (Caplan, 1995; Markon, 2013). Others believe that the descriptive categories of the DSM have prevented the discovery of successful treatments and should be replaced with constructs based on etiology and pathogenesis (Hyman, 2011; Insel, 2013). Still others question the use of diagnostic categories derived from clinical tradition and prefer a more quantitative, dimensional approach to classification (Krueger et al., 2018; Livesley, 2012). In contrast, Parnas and Zandersen deride neither the classification of psychopathology, descriptive psychopathology per se, nor the use of categories/kinds, but instead critique the use of the DSM symptom based, operational approach to classification in the absence of an integrative theoretical framework.

In a commentary on Miller and Bartholomew, Zachar notes that the Research Domain Criteria initiative (RDoC) has always represented an incursion of cognitive neuroscience and factor analytic dimensional models from psychology into biological psychiatry, and that non-reductionist aspirations have been a part of RDoC from the beginning. Rather than reductionism, a more problematic tenet of the RDoC initiative is a contempt for the categories of the DSM, seen in a tendency to blame the lack of progress on adherence to a categorical model rather than on the intrinsic complexity of psychopathology. One aspect of this complexity is the open nature of psychological concepts. Turning to the “neural circuit” metaphor, Zachar suggests that the notion of a “circuit for” may reflect adopting too modular a theory of mind, which as a framework, may unnecessarily constrain a mechanistic science.

Several chapters in this volume address the heterogeneity problem in psychiatric classification, i.e., that constructs such as major depressive disorder and generalized anxiety disorder have turned out to be too coarse-grained to support the discovery of underlying mechanisms or to guide clinicians in selecting disorder-specific treatments (Bilder Chapter 5, Miller Chapter 20, although see Bechtel Chapter 2 for a different view on mechanisms and heterogeneity). Bringing a more statistical perspective to the heterogeneity issue, Jan-Willem Romeijn and Hanna M. van Loo refer to this as the reference class problem.

To define descriptive psychopathology as classification of disorders with respect to manifest signs and symptoms as opposed to deeper causes is a somewhat superficial construal that does not take into account the various ways that something can be brought under a description. The philosopher’s notion of bringing something under a description can be illustrated by a non-behaviorist reading of Gilbert Ryle’s book The Concept of Mind. That things can be brought under more than one description highlights the importance of re-describing. An important example of re-describing psychopathology is the discovery of panic disorder from which five desiderata for useful descriptions and re-descriptions can be derived. With respect to causes, the elucidation of a causal model for a phenotype can often lead us to notice something descriptively that we had not noticed before, in which case the causal model becomes part of a thicker description of the phenotype.

Bill Bechtel is rare breed of philosopher whose writings on mechanistic models are afforded as much respect by scientists in biology and psychology as they are by philosophers. Earlier in his career Bechtel (1988a, 1988b) published two short books written largely for psychologists titled Philosophy of Mind: An Overview for Cognitive Science and Philosophy of Science: An Overview for Cognitive Science. These books demonstrated that he is adept at writing for a cross-disciplinary audience. His work on mechanisms, however, is of a decidedly technical nature, and nevertheless highly admired outside the philosophy of science.

In his chapter, Ken Schaffner dives into the variety of levels that are invoked in both the traditional fear center model and Joe LeDoux and Danny Pine’s (2016) two-system model of fear. In many ways, LeDoux wrote the book on the biology of fear conditioning and for him to claim (with Pine and others) that rodent models can never fully represent the biology of human fear has incredible gravitas.